A R C H I V E S O F M E T A L L U R G Y A N D M A T E R I A L S

Volume 57 2012 Issue 3

DOI: 10.2478/v10172-012-0085-5

J. NIAGAJ∗, Ł. MAZUR∗

FERRITE CONTENT MEASUREMENTS IN S32101 LEAN DUPLEX STAINLESS STEEL AND ITS WELDED JOINTS

POMIARY FERRYTU W STALI LEAN DUPLEX S32101 I JEJ ZŁĄCZACH SPAWANYCH

Due to their mechanical and plastic properties as well as unique corrosion resistance, two-phase lean duplex steels areincreasingly popular in industrial applications e.g. for building waterside fixtures, ships, pipelines or containers. A criticalfactor of welded joints made of such steels is the balance between austenite and ferrite; the latter being measured by a specialdevice called ferritoscope. The article contains the results of tests focused on measurements of ferrite content in S32101 leanduplex steel and its welded joints. The text also presents the impact of such factors as test sample thickness or shape andcondition of measurement surface etc. on test results. In addition, this paper discusses the use of correction factors, describesproblems arising during measurements of ferrite in welded joints and presents manners of the elimination of the latter. Theconducted tests revealed that MAG method welding parameters affect the content of ferrite in butt welded joints produced withS32101 lean duplex steel.

Keywords: lean duplex, welded joint, ferrite measurement, ferrite content, MAG, GMAW

Dwufazowe stale typu lean duplex z uwagi na ich własności wytrzymałościowe i plastyczne oraz szczególną odpornośćkorozyjną są coraz częściej stosowane w przemyśle do budowy armatury nadbrzeżnej, statków, rurociągów lub zbiorników. Wzłączach spawanych z tych stali bardzo ważna jest równowaga pomiędzy austenitem i ferrytem. Pomiary zawartości ferrytuwykonuje się za pomocą np. urządzenia typu ferrytoskop. W artykule przedstawiono wyniki badań zawartości ferrytu zapomocą ferrytoskopu w stali lean duplex S32101 i jej złączach spawanych oraz wpływ na wyniki pomiarów takich czynnikówjak: grubość mierzonej próbki, stan i kształt powierzchni pomiarowej itp. Omówiono stosowanie współczynników korekcyj-nych. Wyszczególniono problemy powstające podczas wykonywania pomiarów ferrytu w złączach spawanych oraz sposoby ichniwelowania. Przeprowadzone badania wykazały, że parametry spawania metodą MAG maja wpływ na zawartość ferrytu wdoczołowych złączach spawanych wykonanych ze stali lean duplex S32101.

1. Introduction

As is commonly known, duplex steels are two-phasestainless steels of ferritic-austenitic structure. The con-tent of each of the phases (α and γ) in the structure ofthese steels is more or less equal and amounts to approx.50%. The qualitative equilibrium between ferrite (phaseα) and austenite (phase γ) in duplex steels contributesto their higher strength and excellent corrosion resis-tance thus combining the advantages of both ferritic andaustenitic stainless steels. [1÷4].

Welding processes, due to heating, melting and so-lidification of both parent and filler metals used for weld-ing duplex steels, change the ferrite-austenite proportionnot only in the weld but also in the heat affected zone(HAZ). Depending on welding conditions and parame-ters as well as the application or failure to a filler metal

(and its content) in the material of the newly formedweld, the ferrite-austenite proportion may change, whichhas its consequences in the modification of mechanicalproperties and corrosion resistance of welded joints. Dueto the foregoing, the ferrite content in the material ofwelded joints made of duplex steels, including standard,super, hyper and lean duplex, should be from 30÷35 to60÷65%, according to various publications the aforesaidrange may be slightly wider or narrower as well as slight-ly shifted upwards or downwards, yet never outside thelimits of the range 20÷70% (ferrite number approx. FN30÷100) [5÷7].

The article presents the results concerning the de-termination of ferrite content in the parent metal andwelded joints produced with lean duplex steel designat-ed as S32101 according to the UNS (Unified Number-ing System). If compared with standard duplex steels,

∗ WELDING TECHNOLOGIES DEPARTMENT, INSTITUTE OF WELDING, 44-100 GLIWICE, 16-18 BŁ. CZESŁAWA STR., POLAND

780

lean duplex steels are characterised by lowered contentof molybdenum and nickel, partially replaced by man-ganese.

2. Experimental procedure

Ferrite content measurements, in the parent met-al were carried out using 6 mm and 12 mm-thick,100×150mm-sized S32101 lean duplex steel samples.In the material of individual zones of welded joints,ferrite content measurements were carried out using12mm-thick, 300×350 mm-sized S32101 lean duplexsteel joints because in this joints high changes in ferritecontent have been measured. The chemical compositionof S32101 lean duplex steel and that of standard du-plex steel (S32205), used for comparative purposes, ispresented in Table 1. The MAG-welded test joints wereproduced with AVESTA LDX 2101 electrode wire (∅1.2mm) of the chemical composition presented in Table2; the shielding gas being Ar + 2.5 % CO2 (designatedas M12-ArC-2.5 according to PN-EN ISO 14175) andthe forming gas being pure argon (designated as I1-Araccording to PN-EN ISO 14175). The test joints were

made using parameters, the selection of which ensuredobtaining proper welded joints with both low and highamount of heat input (Table 3).

3. Results

3.1. Determination of ferrite content with FMP30ferritoscope

The ferrite content in high-alloy stainless steels canbe determined with various methods and measurementdevices [8]. The tests in question involved the use of themost commonly industry-applied Fischer-manufacturedferritoscope FMP30 (Fig. 1) enabling the determinationof ferrite content in steels, deposited metals and welds ofwelded joints having the composition of austenitic steelsand that of austenitic-ferritic duplex steels as well as de-termining martensite content in austenitic steels [9]. Theferrite content can be expressed as a ferrite number (FN)or percentage. The tests were performed with the help ofa measurement probe FGAB1.3-Fe of the measurementrange 0.1-80% Fe (Fig. 1a).

TABLE 1Chemical composition of duplex steel plates applied in research

UNSWerkstoffnumber

Platethickness,

mm

Datasource

Chemical composition, % Ferritecontent,

%C Si Mn Cr Ni Mo Cu N

S32101 1.4162 6 certificate 0.028 0.70 4.90 21.34 1.50 0.19 0.25 0.21 56

S32101 1.4162 12 certificate 0.021 0.66 4.85 21.4 1.64 0.22 0.30 0.22 –S32205/S31803 1.4462 6

PN-EN10088-2 max 0.030 6 1.00 6 2.00 21.0÷23.0 4.5÷6.5 2.50÷3.5 – 0.10÷0.22 –

TABLE 2Chemical composition of applied electrode wire AVESTA LDX 2101

Diameter, mm Data sourceAlloying elements content, %

C Si Mn P S Cr Mo Ni N2 Cu

1.2 certificate 0.016 0.53 0.75 0.029 0.001 23.12 0.25 7.27 0.117 0.17

TABLE 3MAG method welding parameters of test butt joints of lean duplex S32101 steel plates of thickness 12 mm

No.Joint

designation Current intensity, AArc voltage,

VTravel speed,

cm/minHeat input,

kJ/mm

1 D24 122 22.0 28.0 0.50

2 D26 126 22.0 11.1 1.20

3 D27 230 27.6 16.4 1.90

4 D28 224 28.8 6.7 4.60

781

Fig. 1. Ferritoscope FMP30 with measurement probe (a) and ferrite content measurement from face of weld on experimental welding joint (b)

Pursuant to ISO 8249 [10] recommendations, theferrite measurements were made at a minimum of 6points on the surface of each test sample. Due to thefact that the ferrite content in S32101 lean duplex steelexceeds 20 FN, a minimum of 5 readings were madein each of the six measurement points, with the read-ing corresponding to the highest ferrite number valueadopted as the FN measured value [10].

3.2. Ferrite content measurement in parent metal

a) impact of surface condition

It is commonly known [8] that ferrite content mea-surement results obtained with a ferritoscope are con-ditioned on the condition of a measurement surface(roughness, oxidation, presence of paint, varnish, rustor scale) and the thickness of a test sample. In order todetermine the effect of these factors on test results, thetests were performed with 6 mm-thick S32101 steel sam-ples of the surface a) in delivery condition; b) ground;c) painted. Comparative tests were conducted using 6-and 12 mm-thick S32101 steel samples and 6 mm-thick2205 steel samples.

The tests of the 6 mm-thick S32101 steel samplesrevealed the lowest (approx. 30%) ferrite content in caseof the samples with paint-covered surface (Fig. 2). Inturn, the ferrite content revealed on the ground surfaceamounted to 47%, whereas that determined in the sam-ples in delivery condition was only slightly lower andamounted to 45%. A similar dependence could be ob-served in case of 12 mm-thick S32101 steel samples(Fig. 3) and 6 mm-thick 2205 steel samples (Fig. 4).As a result, it is possible to draw a conclusion that thepresence of non-magnetic substances (paint, oxides gen-erated during production and storage of steel etc.) on thesurface of tested objects reduces measurement results.The degree of the aforesaid reduction depends on thethickness of a non-magnetic layer; the presence of paintreduces measurement values by as much as 17% (Fig. 2),whereas the presence of a considerably thinner oxide lay-

er reduces measurement values by not more than 2÷4%if compared with those obtained from a ground surface(Fig. 2÷4).

Fig. 2. Influence of surface condition on ferrite content measurementresults in duplex steel S32101 of 6 mm thickness

Fig. 3. Influence of surface condition on ferrite content measurementresults in duplex steel S32101 of 12 mm thickness

Fig. 4. Influence of surface condition on ferrite content measurementresults in duplex steel S32205 of 6 mm thickness

782

In order to determine the dynamics of decreas-ing measurement value results in the presence of anon-magnetic layer, it was necessary to carry out a test,in which the thickness of the layer was simulated byplacing the ferritoscope measurement probe at a specificheight ”a” over the sample surface within the 0-6 mmrange (Fig. 5). The S32101 steel samples used in testscame from different heat melts characterised by differ-ent ferrite content measured on the surface of plates:A – approx. 48% and B – approx. 37%. In the tests itwas determined that the air space between the tip of themeasurement probe and the measurement surface signif-icantly affects the result of ferrite content measurements(Fig. 6). In case of both samples, the distance of only 0.5mm between their surface and the measurement probedecreased the measurement result by more than a half:from 48% to 22% (sample A) and from 37% to 15%(sample B). If the above results were to be adopted asfinal ones, at least the B-designated steel would have tobe recognised as failing to meet the requirements fromthe ferrite content point of view, although, in fact, it isotherwise. Thus, the presented test results indicate thenecessity of measuring ferrite content in steels but, firstof all, in welded joints, prior to painting (if any) andfollowing the removal of a thick oxide layer, includingthe post-weld temper; this being also indicated in ENISO 15614 series standards regarding the qualificationof welding technologies.

Fig. 5. Scheme of measurement probe location above the test samplesurface

Fig. 6. Correlation between ferritoscope indication and distance ”a”between measurement probe and samples A and B measurement sur-face of ferrite different content

During the tests it was also ascertained that the testsample affects the magnetic field of the measurementprobe when the distance between the two does not ex-ceed 5 mm (Fig. 6), yet the influence at such a consider-able distance is slight and practically insignificant froma = 2÷3 mm onwards.

b) impact of sample thickness

The measurement result is also affected by the thick-ness of a measured element (detail, plate, weld) or ap-plied layer e.g. the thickness of a duplex steel contentsurfacing layer on the surface of unalloyed steel. Figure7b presents the results of measurements carried out onS32101 steel sample of strokly changing thickness from6 to 1 mm (Fig. 7a). The measurements at the step of 1mm thickness are characterised by approx. 7÷12% low-er ferrite content if compared with the sample steps ofthicknesses from 2 to 6 mm. The measurement resultsof the 2÷6 mm samples were between 52 and 56% andare similar to the ferrite content specified in the man-ufacturer’s certificate of the tested S32101 lean duplexsteel i.e. 56% (Table 1).

Fig. 7. Ferrite content measurement in lean duplex S32101 steel sam-ple of strokly changing thickness: a) sample dimensions and measure-ment points location; b) ferrite content measurements results

The change of the thickness of a measured element,similarly as the radius of the curvature of the face orroot reinforcement, can be recognised as the change ofmeasured sample volume. The ferritoscope-based ferritemeasurement in a sample of insufficient volume may leadto obtaining results with errors, which, in turn, might

783

lead to misinterpretation and incorrect conclusions re-lated to the quality of a produced weld or padding weld.The tests made it possible to ascertain that in case ofthe steel samples of 50% ferrite content (duplex steels),approx. 98% of measurement quantity is obtained atthe sample thickness of mere 2÷2.5 mm (Fig. 7). Oneshould, however, bear in mind that the magnetic field ofthe measurement probe influences the test material at thedistance of up to 5 mm (Fig. 6), which could suggest,that ferrite content measurement results closest to realvalues will be those performed in welded joints of theminimum thickness of 4÷5 mm.

On the basis of the above results of ferrite measure-ments conducted on the surface of samples at stroklychanging thickness as well as on grounds of the tests fo-cused on magnetic field influence it is possible to draw aconclusion that in case of S32101 lean duplex steel, cor-rect results of measurements obtained with the FMP30ferritoscope are obtained with the sample of a minimumthickness of 2÷2.5 mm, which, to a significant extent,coincides with the ferritoscope manufacturer’s recom-mendations stating that the minimum thickness of thesample or that of surfacing layer must not be less than2 mm.

c) impact of measurement point location

In addition to the thickness of test elements andnon-magnetic layers covering the surface of test sam-ples, other factors, such as the location of a measure-ment point, also influence ferrite measurement values.Presented below are measurement values of ferrite con-tent on a ground surface and in the transverse section oftwo samples cut out of two S32101 steel plates comingfrom two different heat melts: A and B. During the testsit was determined that the ferrite content measured onthe surface of the samples was approx. 48% in case ofsample A and approx. 37% in case of sample B, whereasin case of the cross section the ferrite content was 54%and 52% respectively (Fig. 8). The difference betweenmeasurement values for individual samples dependingon measurement point location varies and stands at ap-prox. 6% in case of sample A and approx. 15% in caseof sample B. It should also be noted that very similar,and also, in both cases, higher ferrite content values wereobtained during measurements carried out on the trans-verse section surface of the samples: A – approx. 54%and B – approx. 52%. This fact could indicate that boththe quantity of the aforesaid measurement differencesand the higher ferrite content measurement results in thetransverse section of the samples are influenced by theshape of grains and their linear packing resulting fromplate rolling texture (Fig. 9). The magnetic interactionbetween thinner ferrite layers and the ferritoscope mea-

surement probe field differs from the interaction whichcan be observed in case of the absence of laminar pack-ing of grains. In case of the structure characterised by thepresence of thin ferrite layers, lower voltage is inducedin the measurement coil, which, in turn, leads to theunderrating of measurement results [11÷13]. A similarphenomenon can be observed in austenitic-ferritic steels.

Fig. 8. Differences in ferrite measurements results on ground surfaceand in samples cross section area cut out of S32101 steel plates fromdifferent heat melts

Fig. 9. Influence of shape and packing of austenite grains in leanduplex steels from both heat melts on ferrite measurement result onplate surface: A – 48%; B – 37%

The impact of rolling texture on ferrite content mea-surement results was particularly visible when, withinfurther research, ferrite content was measured on thesurface and in the transverse section of welds. The testrevealed that in case of the welds, the differences result-ing from the location of the measurement point (on thesurface or in the cross section of the sample) are veryinsignificant and contained within a 2÷4% range. Theabsence of considerable differences is probably attribut-able to the character of the structure of the welds, whichis characteristic of a cast material characterised by theabsence of clearly directed grains (Fig. 10).

784

Fig. 10. Differences in areas microstructure of lean duplex S32101 steel weld joint of thickness 12 mm: a) parent metal; b) weld (magn.200x)

A significant impact of rolling (grain strain) on themagnetic field between the ferritoscope and the mea-surement probe hinders the execution of proper ferritecontent measurements and limits the possibility of com-paring obtained results among one another in case ofwelded joints made of plates coming from different heatmelts. Ferrite content measurements in the HAZ material(conditioned by the sufficient width of the heat affect-ed zone – min. 2÷3 mm) and that in the parent metalcan be compared only in case of one specific joint. Asa result, one can determine the degree of increasing ordecreasing ferrite content in the HAZ compared with theferrite content determined in the parent metal.

4. Ferrite content measurement in butt weldedjoints made of S32101 steel plates

During the ferrite content measurements conductedon the butt welded joints made of S32101 lean duplexsteel plates (Table 3), ferrite content was measured on thelevel of their face and root following 45-minute chemi-cal etching with the ANTOX 71E paste. Ferrite contentmeasurements were made along the weld axis.

The MAG welding of the test joints made of 12mm-thick S32101 lean duplex steel plates was connect-ed with supplying various heat input (Table 3); the heatinput being within the 0.5÷4.6 kJ/mm range (the steelmanufacturer-recommended heat input being from 0.8to 1.8 kJ/mm [6]. The welded joints made with S32101lean duplex steel using the aforesaid heat input and simi-lar surface condition revealed varied ferrite content (Fig.11 and 12). The ferrite content measurements conduct-ed on the face and root surfaces of the joints producedwith low heat reveal higher values than the correspond-ing measurements carried out in the joints made with

a higher heat input. In case of 0.5 kJ/mm heat supply,the ferrite content in the weld, on the face and root sidewas within 40÷47%, whereas in case of 4.6 kJ/mm heatsupply, the ferrite content was between 32 and 38% (Fig.11 and 12). It should also be mentioned that in all ofthe cases the ferrite content was, in the weld materi-al, slightly higher on the face side and slightly low-er in the root side. The data presented in Figures 11and 12 justify the conclusion that, irrespective of the

Fig. 11. Influence of heat input on ferrite content on weld face in testwelded joints: D24, D26, D27, D28 (Table. 2) lean duplex S32101plates of thickness 12 mm

Fig. 12. Influence of heat input on ferrite content on weld root in testwelded joints: D24, D26, D27, D28 (Table. 2) lean duplex S32101plates of thickness 12 mm

785

location of a measurement point and the heat input sup-plied during MAG welding, the ferrite content in thematerial of the butt welds made of 12 mm-thick S32101lean duplex steel plates was between 32 and 52%.

During the ferrite content measurements performedin welded joints by means of the FMP30 ferritoscope,the most significant problem was caused by a reinforce-ment present both on the face and root side of the weld.The convexity of the reinforcement is one of the cas-es of the change of the volume of material interactingwith the measurement probe magnetic field; this beingcaused by the change of a curvature radius. The impactof the convex curvature radius can be eliminated by theapplication of appropriate correction factors without thenecessity of removing the reinforcement. The values ofthe said coefficients depend not only on the geometry ofsamples or the distance between the measurement pointand the edge of the sample but also on the ferrite contentin an object under investigation. Measurements carriedout on convex surfaces gain, whereas those performed onconcave surfaces lose in value. Curvature radiuses affect-ing ferrite content measurements are smaller in duplexsteels than in case of austenitic-ferritic steels [9].

In order to select a proper correction factor for fer-rite content measurements conducted on the surface offace or root, it is necessary to determine a curvature ra-dius e.g. by measuring the width and height of the faceor root reinforcement. The values of correction factorscan be obtained from appropriate graphs, depending onmeasured ferrite content and the value of a curvatureradius; the aforesaid graphs are usually supplied by thedevice manufacturer usu. in manuals [9]. The selectionof a proper coefficient is rather a time-consuming activ-ity; this being due to changing joint lengths and rein-forcement width and thus changing curvature radius andferrite content values being the basis for the selectionof a proper correction factor. In addition, the reading ofcorrection factors from graphs is subject to inaccuracycaused by insufficient concentration of reference lines forindividual curvature radiuses. The device manufacturerdoes not specify according to what rules one should se-lect the correction factor for diameters, for which thereare no reference lines on the diagram.

During measurements in points of varying width andheight of weld face reinforcement or weld root reinforce-ment one should apply various values of correction fac-tors. Unfortunately, the aforesaid manner of correctioncan be used only after completing measuring actions;this being due to the fact that one can enter (into a mea-surement device) only one corrective factor for a singlemeasurement series [8, 9].

Figure 13 presents ferrite content values measuredon the root surface prior to and following the removal of

the reinforcement. It was determined that the values ofmeasurements performed on the convex surface of theroot were by approx. 10% lower than those obtained af-ter the removal of the reinforcement (after milling). Theunderrated measurement results are caused by the shapeof the surface subject to measurement or, speaking moreprecisely, by too low metal volume affected by the mag-netic field of the measurement probe. The application ofappropriate correction factors caused that the post-weldferrite content measurement conducted on the surface ofthe root was similar to the value of the measurement fol-lowing the removal of the reinforcement (Fig. 13). Thevalues of measurements made on the milled surface andthose with the correction factors applied differ only by2÷3%. Carrying out measurements one should remem-ber that not only the shape of the face and root affects theresults of ferrite content measurement in the weld but al-so the size and arrangement of the shell, which, similarlyas the roughness of the sample surface, cause significantscatter of measurements conducted in one measurementpoint.

Fig. 13. Influence of root weld shape of S32101 steel and correctionfactor on measurement quantity of ferrite content

5. Discussion

The tests focused on the condition and shape ofsample surfaces as well as on the procedure of measure-ment of ferrite content using the FMP30 ferritoscopeequipped with the FGAB1.3-Fe measurement probe inS32101 lean duplex steel and its welded joints revealeda number of peculiarities which should be rememberedwhile working on obtaining measurement results repro-ducing the real content of ferrite in a test material withthe highest possible accuracy.

The tests revealed that the presence of paint on thesurface of samples decreases measurement values by asmuch as 17% i.e. a fall from 47% down to 30% (Fig. 2).In turn, a thin layer of oxides present on the surface ofplates in delivery state decreases the measurement resultif compared with the results obtained from ground sam-ples by mere 2÷4% (Fig. 2÷4), which in case of steelwith 50% ferrite content has no significant impact on the

786

assessment of measurement results. A problem, however,may arise when the content of ferrite in the test steel iswithin the range of allowed boundary values [5÷7], andparticularly around 30%. In the foregoing situation, a de-viation of even a few percent may lead to an indicationthat a given steel does not meet related requirements.The tests simulating a non-magnetic layer on the platesurface revealed that the most significant decrease in ob-tained ferrite content results can be observed at the initialstage of increasing the distance between the measure-ment probe and the sample surface: in case of distance“a” equal only to 0.5 mm, measured ferrite content val-ues decrease by more than a half (Fig. 6). In order to ob-tain accurate and reliable ferrite content data one shouldcarefully prepare the measurement equipment (removeimpurities or paint, or oxide layer) and, if in doubt, con-duct additional tests applying a computer-aided analysisof a microstructural image [8].

One of the limitations accompanying ferrite contentmeasurements with the FMP30 ferritoscope is the thick-ness of a test element or cladded layer of austenitic steelor ferritic-austenitic steel composition produced on thesurface of unalloyed steel. The tests revealed that in caseof S32101 lean duplex steel, proper measurement resultscan be obtained with the sample thickness not less than2÷2.5 mm (Fig. 7). Increasing the thickness of sam-ples above 2÷2.5 mm leads to only a slight increase ofobtained results, not exceeding 3÷4%. The most appro-priate ferrite content measurement results are, however,obtained with the sample thickness of at least 4÷5 mmbecause, as it was previously determined, this is the dis-tance of the magnetic field effect of the measurementprobe applied (Fig. 6).

Moreover, the tests also revealed that, in additionto the condition of the surface and its thickness, alsothe position of the measurement point plays an impor-tant role. It was determined that in case of plates madeof S32101 lean duplex steel coming from two differentheat melts, the ferrite content measured in the transversesection was higher than that determined on the surfaceof the samples: approx. 54 and 48% in case of sampleA, and 52 and 37% in case of sample B respectively(Fig. 8). As can be seen, in case of steel B the differencebetween measurement values amounts to approx. 15%.Such a significant difference could, in extreme cases,prevent further processing of a given steel plate. Themetallographic examination showed that the microstruc-ture of steels A and B differs significantly just belowthe surface of the plates (Fig. 9). Steel B has the struc-ture characterised by grains elongated in the directionof rolling, which is highly possible to be responsible forinducing lower current in the measurement probe, which,in turn, translates to a lower ferrite content measurement

result. The aforesaid problems, however, do not occur incase of welded joints as welds are characterised by thestructure typical of cast materials, in which, one does notusually observe clear elongation of grains depending onthe position of the microsection plane (Fig. 10). Differ-ences resulting from the place of measurement (on thesurface or in the transverse section of welds) are usuallycontained within a 2÷4% range.

The primary purpose of the tests in question wasto measure ferrite content in MAG-welded joints madeof S32101 lean duplex steel. The tests revealed that theferrite content measured both on the face and root sur-faces of the weld made with a low heat input (approx.0.5 kJ/mm) is between 40 and 47% and is higher incomparison with the results obtained with a relativelyhigh heat input (approx. 4.6 kJ-mm), being between 32and 38% (Fig. 11 and 12). It should me mentioned thatin all of the test joints, ferrite content was slightly higherin the weld material on the face side than it was on theroot side. According to the data presented in Figures 11and 12, in case of MAG-method welding, irrespective ofthe place of measurement and the heat input in 0.5÷4.6kJ/mm range, the ferrite content in the material of buttwelds made of 12 mm-thick S32101 lean duplex steelplates is between 32 and 52% and thus is containedwithin the recommended range of values from 30÷35to 60÷65% [5÷7].

During the ferrite content measurements performedin welded joints by means of the FMP30 ferritoscope,the most significant problem was caused by a reinforce-ment both on the face and root side of the weld, whichthrough the change of the volume of the material inter-acting with the magnetic field of the measurement probeled to the underrating of measurement values. The im-pact of the convex curvature radius was eliminated eitherby milling or through the application of appropriate cor-rection factors [9] without the necessity of removing thereinforcement. During the tests it was determined (Fig.13) that the values of the measurements performed onthe convex surface of the root were by approx. 10% lowerthan those obtained after the removal of the reinforce-ment (after milling). The application of appropriate cor-rection factors caused that the post-weld ferrite contentmeasurement conducted on the surface of the root wassimilar to the value of the measurement following the re-moval of the reinforcement. The values of measurementsmade on the milled surface and those with the correc-tion factors applied differ only in 2÷3%. It should alsobe mentioned that while carrying out measurements onthe surface of the face or that of the root one should alsotake into account the size and arrangement of the shell,which, similarly as the roughness of the sample surface,

787

cause significant scatter of measurements conducted inone measurement point.

6. Conclusions

As a result of the tests focused on ferrite contentin S32101 lean duplex steel and its welded joints, per-formed by means of the FMP30 ferritoscope with theFGAB1.3-Fe measurement probe it was established that:1. Proper preparation of a surface to be measured con-

sisting in the removal of impurities, oxide layer,post-weld temper or paint as well as the applica-tion of proper correction factors while carrying outferrite content measurements on the convex surfaceof the face or root of the weld enable obtaining mea-surement values which are very close to the real fer-rite content measured in S32101 lean duplex steeland its welded joints; the differences with the actualstate being not more than 2÷4%. The presence ofpaint on the measurement surface of failure to ap-ply correction factors increases differences betweenmeasurement values by as much as 20%, which, insome cases, unacceptably distorts the actual ferritecontent in test steel samples or welds.

2. Depending on a heat melt, ferrite content measure-ment values on the surface of the plate and in itstransverse section may vary by up to approx. 15%,which is attributable to the packing of grains char-acteristic of rolling. In case of welds, the said dif-ferences amount to mere 2÷4%, which is the conse-quence of the formation of a structure typical of acast material

3. During welding with the MAG method, irrespectiveof the place of measurement and the changing heatinput varying between 0.5 and 4.6 kJ/mm, the fer-rite content in the material of butt welds made of

12 mm-thick S32101 lean duplex steel plates is be-tween 32 and 52% and is contained within the al-lowed range.

REFERENCES

[1] R.N. G u n n, Duplex stainless steel: microstructureproperties and applications. Woodheat Publishing Ltd,England (1997).

[2] R. B a d j i a, M. B o u a b d a l l a h b, B.B a c r o i x a, C. K a h l o u n a, K. B e t t a h a r c, N.K h e r r o u b a b, Materials Science and EngineeringA 496, 447-454 (2008).

[3] J.W. E l m e r, T.W. E a g a r, Welding Journal 4,141-s-150-s (1990).

[4] A. B a c z m a ń s k i at al., Archives of Metallurgy andMaterials 53, 1, 89-96 (2008).

[5] Join publication, How to weld Duplex Stainless Steel.Avesta Welding (2006).

[6] Informative materials, Fabricating LDX 2101 R©. RolledAlloys (2009).

[7] NACE Standard MR0103-2007, Materials Resistant toSulfide Stress Cracking in Corrosive Petroleum RefiningEnvironments.

[8] J. N i a g a j, Ł. M a z u r, Bulletin of Polish WeldingInstitute 4, 49-59 (2010) (in Polish).

[9] Sernice manual of ferrytscope FMP 30. Helmut FischerGMBH (2008).

[10] ISO 8249:2000 Welding – Determination of FerriteNumber (FN) in austenitic and duplex ferritic-austeniticCr-Ni stainless steel weld metals (in Polish).

[11] D. J i l e s, Introduction to magnetism and magnetic ma-terials. Taylor & Francis Group, (1998).

[12] Joint publication, Handbook of Magnetism and Ad-vanced Materials, John Wiley & Sons Ltd, England 1(2007).

[13] Joint publication, Handbook of Magnetism and Ad-vanced Materials, John Wiley & Sons Ltd, England 2(2007).

Received: 10 February 2012.